Shield tunnel section fitting method, slab staggering quantity detection method and detection system

By randomly extracting a small amount of point cloud data to fit an elliptical curve in the shield tunnel cross-section fitting method and using deviation weights to evaluate the reliability of the fitting results, combined with Hough transform to locate the annular joint, the problems of low efficiency and reliance on manual labor in traditional detection methods are solved, and efficient and accurate shield tunnel misalignment detection is achieved.

CN119043268AInactive Publication Date: 2024-11-29CHINA RAILWAY SOUTHWEST SCI RES INST CO LTD

Patent Information

Application Number
CN202411534048.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-31
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional methods for detecting misalignment in shield tunnels are inefficient and rely on manual experience, failing to meet the demands for high-efficiency detection. Furthermore, existing 3D laser scanning technology requires cumbersome preprocessing steps during the fitting process, impacting detection efficiency and accuracy.

Method used

The shield tunnel cross-section fitting method is adopted. By randomly extracting a small amount of point cloud data from the three-dimensional cross-section point cloud data, an elliptical curve is fitted. The reliability of the fitting result is evaluated by using deviation weights. Noise points are automatically filtered out and the fitting is performed directly without simplification, filtering or segmentation steps. The Hough transform method is combined to locate the annular joint and calculate the misalignment.

Benefits of technology

It achieves efficient and accurate shield tunnel cross-section fitting and misalignment detection, reduces manual intervention, improves detection efficiency and accuracy, meets specification requirements, and is suitable for actual engineering detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119043268A_ABST
    Figure CN119043268A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of tunnel engineering measurement, in particular to a shield tunnel section fitting method, a slab staggering quantity detection method and a detection system. The shield tunnel section fitting method comprises the following steps: a, acquiring three-dimensional section point cloud data of a shield tunnel; b, randomly extracting at least five pieces of three-dimensional section point cloud data from the three-dimensional section point cloud data, and fitting an elliptic curve based on the extracted three-dimensional section point cloud data; c, the deviation weight of each three-dimensional section point cloud relative to the elliptic curve is calculated, the number of first point clouds with the deviation weights meeting a first threshold value is counted, and the deviation weights are used for describing the deviation degree of each section point cloud relative to the fitted elliptic curve; d, if the number of the first point clouds meets a second threshold value, outputting an elliptic curve; otherwise, the method returns to the step b for iterative execution. The method can perform section contour fitting for the original three-dimensional section point cloud, and is suitable for actual detection engineering.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of tunnel engineering measurement, and in particular to a shield tunnel section fitting method, a misalignment detection method and a detection system. Background Art

[0002] Segment misalignment in shield tunnels primarily refers to the relative displacement between adjacent segments along the tunnel's longitudinal axis. The main causes of segment misalignment include: improper segment assembly according to specifications, uneven jacking pressure, improper grouting pressure control, bolt relaxation, and the complex external environment of the tunnel. This misalignment often leads to defects such as localized segment damage (cracking), tunnel water seepage, grout leakage from the shield tail, and segment loss. These defects threaten traffic safety during tunnel operation.

[0003] Traditional methods for detecting misalignment in shield tunnels include manual inspections and total station scanning. Manual inspections involve directly measuring visible misalignment with a steel ruler. Manual inspections are typically limited to measuring the sides, making it difficult to measure the crown. Total station scanning requires measurement during downtime, resulting in low detection efficiency and failing to meet the demands of the booming subway tunnel inspection market. Traditional misalignment detection methods consume significant manpower and resources, occupying excessive time during operational maintenance windows, resulting in low detection efficiency and a dependence on the accuracy of the results on the experience of the inspectors.

[0004] 3D laser scanning technology offers the advantages of non-contact measurement, high precision and resolution, and rich information for acquiring spatial information. Compared to traditional measurement methods such as total stations and profilers, 3D laser scanning can capture point cloud data without requiring frequent observation point setup. It also offers high data density, high quality, and precision, with measurement accuracy reaching millimeter levels. Therefore, 3D laser scanning is commonly used for detecting cross-sectional deformation, detecting defects such as water seepage and block fall in tunnels, and detecting cracks.

[0005] Shield tunnel cross-sections are typically fitted directly using the least squares method. However, this method requires fitting all data objects, and local noise significantly interferes with the fitting results. Therefore, when directly using the least squares method to fit the cross-section ellipse, the cross-section point cloud data must be noise-filtered to ensure that the fitting results meet the requirements. In actual engineering, however, the steps of cross-section point cloud simplification, filtering, and segmentation are cumbersome, and in order to fit the cross-section of a real-world shield tunnel, it is more desirable to directly fit the original cross-section point cloud data. Summary of the Invention

[0006] The present application provides a shield tunnel section fitting method, a misalignment detection method and a detection system, which directly fit a small number of original section point cloud data, and use deviation weights to evaluate the usability of the fitting results during the fitting process to automatically filter out noise points during the fitting process, eliminating the need for streamlining, filtering, segmenting and other steps of the section point cloud data before fitting, making it suitable for actual detection projects.

[0007] This application is implemented through the following technical solutions: In a first aspect, the present application provides a shield tunnel cross-section fitting method, comprising the following steps: a. Obtain 3D cross-section point cloud data of the shield tunnel; b. randomly extracting at least five three-dimensional cross-sectional point cloud data from the three-dimensional cross-sectional point cloud data and fitting an elliptic curve based on the extracted three-dimensional cross-sectional point cloud data; c. Calculating a deviation weight of each three-dimensional cross-section point cloud relative to the elliptic curve, and counting the number of first point clouds whose deviation weights satisfy a first threshold, wherein the deviation weights are used to describe the degree of deviation of each cross-section point cloud relative to the fitted elliptic curve; d. If the number of the first point cloud meets the second threshold, output the elliptic curve; otherwise, the method returns to step b and iterates.

[0008] The shield tunnel section fitting method provided in this application selects a small number of point cloud data from three-dimensional section point cloud data (original section point cloud data) for elliptic curve fitting; the method evaluates the degree of deviation between each three-dimensional section point cloud and the fitted elliptic curve by calculating the deviation weight, thereby realizing automatic filtering of noise points; by counting the number of first point clouds, the method further evaluates the reliability of the elliptic curve to ensure that the fitting error meets the specification requirements; this method is particularly suitable for actual detection projects, and can directly process three-dimensional section point cloud data without going through preprocessing steps such as simplification, filtering or segmentation.

[0009] In some optional embodiments, the deviation weight is obtained by: Solve the coordinates of the two foci and the length of the major axis of the elliptic curve; Calculate the sum of the distances between the shield tunnel cross-section point cloud and the two focal points; The absolute value of the major axis length, spacing and difference is used as the deviation weight.

[0010] In some optional embodiments, step d further comprises: Before outputting the elliptic curve, execute steps b and c again to obtain the second point cloud number; If the number of the second point cloud is greater than that of the first point cloud, the elliptic curve is updated.

[0011] In some optional embodiments, step b further comprises: Mark the two intersection points of the tunnel profile with the tunnel floor; With the shield tunnel's central axis as the center, construct the major arc between the two intersection points; Mark the points dividing the major arc into five equal parts and construct rays from the central axis to the points dividing into five equal parts to obtain five sector-shaped areas; Randomly select a 3D cross-section point cloud data in each of the five sector areas; The least squares method is used to perform ellipse fitting on the selected three-dimensional cross-section point cloud data.

[0012] In a second aspect, the present application provides a method for detecting the amount of misalignment in a shield tunnel section, which is implemented based on any of the shield tunnel section fitting methods described in the first aspect, and includes the following steps: e. Obtaining the three-dimensional point cloud data of the shield tunnel and obtaining the annular seam position based on it; f. Along the extension direction of the shield tunnel, the first section and the second section are obtained before and after the annular seam; g. Using a shield tunnel section fitting method to obtain the first section profile and the second section profile of the first section and the second section respectively; h. Fit the first section profile and the second section profile and calculate the misalignment.

[0013] In some optional embodiments, step e includes: Map the shield tunnel 3D point cloud data into a 2D grayscale image; Locate the annular seam contour through two-dimensional grayscale images; The Hough transform method is used to fit the annular seam and filter out noise points; The position of the annular seam is obtained based on the mapping relationship between the three-dimensional point cloud data and the two-dimensional grayscale image pixels.

[0014] In some optional embodiments, mapping the shield tunnel three-dimensional point cloud data into a two-dimensional grayscale image includes: The plane point cloud is rasterized using a 3~8mm grid.

[0015] In some optional embodiments, the planar point cloud is rasterized using a 5 mm grid.

[0016] In some optional embodiments, the misalignment amount is calculated by: Calculate the distance r1 from point m on the first cross-section profile to the center of the elliptical curve when the inclination angle is α: r1=a1*b1 / [a1 2 *(sin(θ1-α)) 2 +b1 2 *(cos(θ1-α)) 2] 1 / 2 ; Wherein, a1 is the length of the major semi-axis of the elliptic curve corresponding to the first cross-sectional profile, b1 is the length of the minor semi-axis of the elliptic curve corresponding to the first cross-sectional profile, and θ1 is the angle between the ray extending from the center of the elliptic curve corresponding to the first cross-sectional profile to point m and the horizontal plane; Calculate the distance r2 from point n on the second cross-section profile to the center of the elliptical curve when the inclination angle is β: r2=a2*b2 / [a2 2 *(sin(θ2-β)) 2 +b2 2 *(cos(θ2-β)) 2 ] 1 / 2 ; Wherein, a2 is the length of the major semi-axis of the elliptic curve corresponding to the second cross-sectional profile, b2 is the length of the minor semi-axis of the elliptic curve corresponding to the second cross-sectional profile, and θ2 is the angle between the ray extending from the center of the elliptic curve corresponding to the second cross-sectional profile to point n and the horizontal plane; Calculate the misalignment Δr: Δr=|r1-r2|; Among them, in the extension direction of the shield tunnel, point m and point n coincide.

[0017] In some optional embodiments, the distance between the first section and the second section and the annular seam is 6 cm.

[0018] In a third aspect, the present application provides a shield tunnel section misalignment detection system, comprising a three-dimensional laser scanning device and a computer device; The three-dimensional laser scanning device is used to scan the shield tunnel to obtain three-dimensional point cloud data of the shield tunnel, and is used to send the three-dimensional point cloud data to the computer device; The computer device includes a storage module, a processor, and a first computer program and a second computer program stored on the storage module and executable on the processor, wherein when the processor executes the first computer program, any one of the shield tunnel section fitting methods described in the first aspect is implemented, and when the processor executes the second computer program, any one of the shield tunnel section misalignment detection methods described in the second aspect is implemented.

[0019] Compared with the prior art, this application has the following advantages and beneficial effects: The shield tunnel section fitting method, misalignment detection method and detection system provided in this application select a small number of point cloud data from three-dimensional section point cloud data (original section point cloud data) for elliptical curve fitting; the method evaluates the degree of deviation between each three-dimensional section point cloud and the fitted elliptical curve by calculating the deviation weight, thereby realizing automatic filtering of noise points; by counting the number of first point clouds, the method further evaluates the reliability of the elliptical curve to ensure that the fitting error meets the specification requirements; the method is particularly suitable for actual detection projects, and can directly process three-dimensional section point cloud data without going through preprocessing steps such as simplification, filtering or segmentation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present application, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present application and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings: Figure 1 A schematic flow chart of a shield tunnel cross-section fitting method provided in an embodiment of the present application; Figure 2 A schematic flow chart of a method for detecting cross-section misalignment of a shield tunnel provided in an embodiment of the present application; Figure 3 A schematic diagram of the three-dimensional laser scanning principle provided in an embodiment of the present application; Figure 4 Schematic diagram of the process of mapping a three-dimensional point cloud into a two-dimensional point cloud provided in the embodiment of the present application; Figure 5 Schematic diagram of the annular seam extraction process provided in an embodiment of the present application; Figure 6 A schematic diagram comparing the fitting results obtained by directly fitting a 3D cross-section point cloud using the least squares method and directly fitting a 3D cross-section point cloud using a shield tunnel cross-section fitting method provided in an embodiment of the present application; Figure 7 Schematic diagram of the cross-section extraction process on both sides of the annular seam provided in the embodiment of the present application; Figure 8 A schematic diagram of the results of fitting the cross sections on both sides of the annular seam provided in an embodiment of the present application; Figure 9 A schematic diagram of the statistical results of the 0-360 degree misalignment of a single loop provided in an embodiment of the present application; Figure 10 A schematic diagram of the original three-dimensional point cloud data of a shield tunnel provided in an embodiment of the present application; Figure 11 Schematic diagram of the maximum misalignment distribution between rings provided in the embodiment of this application. DETAILED DESCRIPTION

[0021] In order to make the objectives, technical solutions and advantages of this application more clear, the present application is further described in detail below in conjunction with examples and drawings. The schematic implementation methods of this application and their descriptions are only used to explain this application and are not intended to limit this application.

[0022] In the first aspect, the present application provides a shield tunnel cross-section fitting method, such as Figure 1 As shown in FIG, the shield tunnel section fitting method includes the following steps: Step S10: Acquire three-dimensional cross-sectional point cloud data of the shield tunnel.

[0023] In the embodiment of the present application, the three-dimensional cross-sectional point cloud data is the original cross-sectional point cloud data on the shield tunnel cross-section.

[0024] Step S11: randomly extract at least 5 three-dimensional cross-sectional point cloud data from the three-dimensional cross-sectional point cloud data and fit an elliptic curve based on the extracted three-dimensional cross-sectional point cloud data.

[0025] In the embodiment of the present application, when the elliptic curve is expressed by an elliptic equation, it has more parameters. Therefore, when selecting three-dimensional cross-sectional point cloud data, the number thereof is at least 5. In actual implementation, the number of randomly extracted three-dimensional cross-sectional point clouds can be configured to 5. In other embodiments, the number of three-dimensional cross-sectional point clouds can also be configured to 6, 7, or 8. Of course, the number of three-dimensional cross-sectional point clouds should not be too many, as this will increase the probability of extracting noise points. Configuring the number of three-dimensional cross-sectional point clouds to 5 can ensure a considerable fitting speed on the basis of achieving elliptical fitting.

[0026] When fitting elliptic curves with three-dimensional cross-section point cloud data, methods such as least squares method, direct method, Hough transform method, genetic algorithm, RANSAC method, etc. can be used for fitting. Considering the difficulty, computational efficiency, robustness and other performance, in actual implementation, the least squares method can be used to fit elliptic curves with five three-dimensional cross-section point cloud data. Specifically, the general equation of the ellipse can be expressed by formula (1): Ax 2 +Bxy+2Cy+Dx+Ey+1=0 Formula (1); After simplification, we can get formula (2): NP=L formula (2); Where P = [A, B, C, D] T ,L=[-1,-1,……,-1] T ,M=[x 2 ,xy,y 2 ,x,y] T ,N=[M1,M2,……,Mn ] T .

[0027] The least squares method is used to solve equation (3): P=(N T N) -1 N T L formula (3).

[0028] According to the ellipse fitting results, the coordinates of the center of the ellipse (x c ,y c ): Formula (4); Formula (5).

[0029] According to equations (6) and (7), the length of the major axis a and the length of the minor axis b of the ellipse are calculated as follows: Formula (6); Formula (7).

[0030] In order to reduce the probability that the randomly selected 5 3D cross-section point cloud data are all noise points, the following method can be used to randomly select 3D cross-section point cloud data: Mark the two intersection points of the tunnel profile with the tunnel floor.

[0031] The two intersection points can be determined by the design parameters of the shield tunnel.

[0032] With the shield tunnel's central axis as the center, construct the major arc between the two intersection points.

[0033] Mark the points that divide the major arc into five equal parts, and construct rays from the median axis to the points that divide it into five equal parts to obtain five sector-shaped areas.

[0034] A 3D cross-section point cloud data is randomly selected from each of the five sector areas.

[0035] The least squares method is used to perform ellipse fitting on the selected three-dimensional cross-section point cloud data.

[0036] Through the above method, the probability that the five three-dimensional cross-sectional point cloud data are all noise points can be reduced, that is, the distribution of the five three-dimensional cross-sectional point cloud data on the tunnel cross-section is relatively scattered, and the probability that the selected three-dimensional cross-sectional point cloud data meets expectations is greatly improved. This can prevent the selected five three-dimensional cross-sectional point cloud data from always concentrating in areas with more noise points, thereby achieving the purpose of reducing the number of ellipse fitting times.

[0037] S12. Calculate the deviation weight of each three-dimensional cross-section point cloud relative to the elliptic curve, and count the number of first point clouds whose deviation weights meet a first threshold, wherein the deviation weights are used to describe the degree of deviation of each cross-section point cloud relative to the fitted elliptic curve.

[0038] In the embodiment of the present application, the larger the deviation weight, the farther the three-dimensional cross-sectional point cloud is from the fitted elliptical curve. For example, if the elliptical curve is located on the solid cross-sectional surface of the shield tunnel, the larger the deviation weight of the three-dimensional cross-sectional point cloud of the tracks, wires, etc. located in the tunnel will be. These noise points can be regarded as noise points. These noise points are filtered out by the first threshold. If the number of remaining three-dimensional cross-sectional point clouds is large, it indicates that the fitted elliptical curve is closer to the cross-sectional contour of the shield tunnel.

[0039] The deviation weight can be characterized by its position relative to the elliptic curve. Specifically, when the 3D cross-section point cloud is on the elliptic curve, the sum of the distances between the 3D cross-section point cloud and the two foci of the elliptic curve is equal to the major axis length. If the 3D cross-section point cloud is on both sides of the elliptic curve, the sum of the distances between the 3D cross-section point cloud and the major axis length will differ. This difference is the deviation weight, which can be specifically expressed as Δl=||QF1|+ |QF2| - 2 r 1|, where F1 and F2 are the two foci of the fitted elliptic curve, Q is a 3D cross-section point cloud on the shield tunnel section, and r1 is the length of the major semi-axis. If the deviation weight of the 3D cross-section point cloud is less than the first threshold, the count is increased by 1. The deviation weights of all 3D cross-section point clouds are calculated, and the total number of 3D cross-section point clouds with deviation weights less than the first threshold is counted.

[0040] S13. If the number of the first point cloud meets the second threshold, output the elliptic curve; otherwise, the method returns to step S11 and iterates.

[0041] As mentioned above, the more the number of first point clouds is, the closer the fitted elliptic curve is to the cross-sectional contour of the shield tunnel. In specific implementation, a second threshold can be preset. When the number of first point clouds is greater than the second threshold, the elliptic curve can be evaluated as an elliptic curve that meets expectations. If the number of first point clouds is not greater than the second threshold, 5 three-dimensional cross-sectional point clouds are randomly selected again, and the elliptic curve is refitted, and then the elliptic curve is evaluated based on the number of first point clouds.

[0042] During the actual detection process, the second threshold value may be 95%, 90%, 85%, etc. of the total number of the three-dimensional cross-section point cloud. Specifically, different second threshold values ​​may be preset according to different shield tunnels, different detection conditions, etc.

[0043] The shield tunnel section fitting method provided in this application selects a small number of point cloud data from three-dimensional section point cloud data (i.e., original section point cloud data) for elliptic curve fitting; the method evaluates the degree of deviation between each three-dimensional section point cloud and the fitted elliptic curve by calculating the deviation weight, thereby realizing automatic filtering of noise points; by counting the number of first point clouds, the method further evaluates the reliability of the elliptic curve to ensure that the fitting error meets the specification requirements; the method is particularly suitable for actual detection projects, and can directly process three-dimensional section point cloud data without going through pre-processing steps such as simplification, filtering or segmentation.

[0044] During the elliptic curve fitting process, when the number of the first point clouds meets the second threshold, the fitted elliptic curve is not necessarily the best quality elliptic curve. For example, when five different three-dimensional cross-sectional point cloud data are randomly selected, the number of first point clouds is greater. Therefore, in some optional embodiments, when the number of the first point clouds meets the second threshold, five three-dimensional cross-sectional point clouds are randomly selected again for elliptic curve fitting, and the number of the first point clouds is counted. That is, step S13 further includes: S131 . Before outputting the elliptic curve, execute steps S11 and S12 again to obtain a second number of point clouds.

[0045] S132: If the number of the second point cloud is greater than the number of the first point cloud, update the elliptic curve.

[0046] In the embodiment of the present application, when the number of the second point cloud is greater than the number of the first point cloud, it indicates that the reliability of the elliptic curve obtained by the subsequent fitting is higher than that of the previous one. This method can increase the probability of obtaining a higher-quality elliptic curve. In other embodiments, when the number of the first point cloud meets the second threshold, if conditions permit, such as sufficient detection time or higher detection accuracy requirements, 3D cross-sectional point clouds can be randomly selected multiple times and ellipse fitting can be performed. If the number of the second point cloud obtained is greater than the number of the first point cloud, the elliptic curve is updated.

[0047] Compared with the conventional method of directly fitting all original three-dimensional cross-section point cloud data using the least squares method, the feasibility verification of the shield tunnel cross-section fitting method provided in the embodiment of the present application is provided below.

[0048] The fitting results can be found in Figure 6 And Table 1, among which, Figure 6 (a) in the figure is the original 3D cross-section point cloud. Figure 6 (b) is the difference diagram between the elliptic curve and the cross-section contour after directly fitting all the original three-dimensional cross-section point cloud data using the least squares method. Figure 6 (c) is a diagram showing the difference between the elliptic curve and the cross-section profile after fitting using the shield tunnel cross-section fitting method provided in the embodiment of the present application; wherein, Figure 6(a) in Figure 6 (b) and Figure 6 The scatter plot within the smooth edge contour in (c) represents discrete noise points in the tunnel, such as wires, pipes, and tracks. Figure 6 As can be seen from (b) in the figure, the curve fitted by the least squares method deviates from the cross-sectional profile of the tunnel, the fitting result is not ideal, and the fitting error is large. From Table 1, it can be seen that the fitting errors of the major and minor axes both exceed the 10mm specified in the specification, and the ellipticity is 8.5%, which exceeds the 0.6% specified in the specification. Therefore, the effect of directly fitting the original point cloud section using the least squares algorithm is not good. However, the fitting errors and ellipticity of the major and minor axes of the shield tunnel section fitting method provided in the embodiment of the present application meet the requirements of the specification. After 6 iterations, the fitting result can be obtained in only 0.061s of calculation. It can be seen that the shield tunnel section fitting method provided in the embodiment of the present application has good anti-interference ability. In summary, the shield tunnel section fitting method provided in the embodiment of the present application has good adaptability to the original three-dimensional section point cloud fitting.

[0049] Table 1 Original cross-section point cloud fitting results Fitting method Fitting semi-major axis (m) Fitting semi-minor axis (m) Semi-major axis fitting error (mm) Minor semi-axis fitting error (mm) Ovality (%) Number of fitting iterations Run time (ms) Least Squares Method 2.6885 2.4593 11.5 240.7 8.5 1 1 Shield tunnel cross-section fitting method 2.7070 2.7026 7 3 0.16 6 61 In a second aspect, the present application provides a method for detecting the misalignment of a shield tunnel section, such as Figure 2 As shown, the shield tunnel section misalignment detection method includes the following steps: S20. Acquire the three-dimensional point cloud data of the shield tunnel and obtain the annular seam position based on the three-dimensional point cloud data.

[0050] In the embodiment of the present application, 3D point cloud data can be obtained using a 3D laser scanning device. In actual implementation, the GRP5000 measurement system of Amberg can be used to scan the subway shield tunnel. The GRP5000 measurement system integrates a 3D laser scanner, a CCD camera, a self-propelled vehicle, a computer, etc., and is currently the most widely used measurement system in the tunnel inspection market. When the self-propelled vehicle moves on the track, the 3D laser scanner emits a laser beam while rotating at high speed. The laser beam performs a high-density scan of the tunnel surface at the millimeter level to obtain the intensity information of the tunnel point cloud. At the same time, the CCD camera is used to take high-definition photos of the tunnel surface to obtain the color information of the tunnel point cloud, such as Figure 3 As shown, the coordinate axis direction is usually the forward direction of the self-propelled vehicle as the y-axis, the horizontal direction as the x-axis, and the height direction as the z-axis.

[0051] After obtaining the 3D point cloud data, the annular seam can be identified based on the 3D point cloud data. Specifically, the following methods can be used: S201. Map the three-dimensional point cloud data of the shield tunnel into a two-dimensional grayscale image.

[0052] like Figure 4As shown, Figure 4 (a) in the figure is the original 3D point cloud. Figure 4 (b) is the 3D point cloud after segmentation. Figure 4 (c) is based on Figure 4 (b) The expanded two-dimensional point cloud image, Figure 4 (d) is a schematic diagram of rasterizing a two-dimensional point cloud. Figure 4 (e) in Figure 4 (c) in the figure is a two-dimensional grayscale image obtained after rasterization. The three-dimensional point cloud data is first rotated to make the central axis of the tunnel parallel to the absolute coordinate axis. Then, filtering, denoising, and segmentation are performed to remove the tracks, bolt holes, pipes, and cables outside the segments, and the coordinates of the ellipse center are calculated. At this time, the Cartesian coordinate system (x, y, z) corresponding to the three-dimensional point cloud data is converted to the polar coordinate system (ρ, y, θ). Since tunnels often deform after being subjected to external forces and are no longer standard circles but ellipses, the segmented three-dimensional point cloud data is projected onto a standard cylindrical section. At this time, the coordinate system is converted to the cylindrical coordinate system (R, y, θ) and unfolded into a two-dimensional point cloud image at the top of the cylinder. The two-dimensional point cloud is rasterized to obtain a two-dimensional grayscale image in image format.

[0053] A grayscale image is essentially a two-dimensional matrix composed of pixel values. The theoretical sampling spacing for a shield tunnel's three-dimensional point cloud is 2mm. In reality, due to the vibration of the track trolley, the actual sampling spacing is greater than 2mm. Experimental analysis shows that using a grid size of less than 3mm to rasterize a planar point cloud yields clear point clouds, but interpolation results in the appearance of many non-existent points, which appear as black dots in the grayscale image, affecting overall quality. Using a grid size of 8mm or larger results in unclear point clouds and fuzzy grayscale images, making them unsuitable for subsequent processing. Based on comprehensive considerations, a grid size of 3-8mm is used for rasterizing two-dimensional point clouds. In practice, a grid size of 5mm is used for rasterization. The point cloud's intensity is mapped to a range of 0-255, and the average intensity of each grid point is used as the grayscale value of that grid pixel. This results in a grayscale image of good overall quality. Through this point cloud processing, a two-dimensional grayscale image with a resolution of 5mm per pixel is ultimately obtained.

[0054] S202: Locate the annular seam contour using a two-dimensional grayscale image.

[0055] like Figure 5 As shown, Figure 5 (a) in the figure is a two-dimensional grayscale image. Figure 5(b) is a two-dimensional grayscale image after filtering and binarization. In actual implementation, image processing methods can be used to filter the two-dimensional grayscale image through a custom filter, and then image binarization and Canny operator edge extraction can be performed to find the outline of the annular seam.

[0056] S203, using the Hough transform method to fit the annular seam and filter out noise points.

[0057] The ring seam is fitted using Hough transform line detection, such as Figure 5 As shown, Figure 5 (c) in the figure is a two-dimensional grayscale image obtained after Hough transform. Figure 5 (d) is a schematic diagram obtained by filtering out noise after Hough transform. Figure 5 The red vertical line in (c) represents the fitted straight line, and then the noise is filtered out by image filtering methods such as binarization, expansion, and corrosion, and the final result is Figure 5 The black straight line in (d) represents the location of the annular gap.

[0058] S204 , obtaining the annular seam position according to the mapping relationship between the three-dimensional point cloud data and the two-dimensional grayscale image pixels.

[0059] The mapping relationship between 3D point cloud data and 2D image pixels automatically determines the location of the annular seam. This location is determined based on the average of the pixels to the left and right of the black line. Practical verification shows that the errors between the five locations corresponding to the five processed annular seams and the actual mileage are 2mm, 2mm, 4mm, 0mm, and 3mm, respectively, with a maximum error of 4mm. This recognition is end-to-end, with no cumulative error. According to relevant literature, slices are generally taken from a maximum of 10cm on either side of the annular seam to determine the amount of misalignment. Therefore, a 4mm error is almost negligible for misalignment analysis.

[0060] S21. Along the extension direction of the shield tunnel, the first section and the second section are obtained before and after the annular seam.

[0061] In actual implementation, the distances between the first section and the second section and the annular seam can be 6 cm respectively.

[0062] S22. Use any one of the shield tunnel section fitting methods provided in the first aspect to obtain the first section profile and the second section profile of the first section and the second section respectively.

[0063] S23, fitting the first cross-sectional profile and the second cross-sectional profile, and calculating the misalignment amount.

[0064] The amount of misalignment can be calculated using the following: Calculate the distance r1 from point m on the first cross-section profile to the center of the elliptical curve when the inclination angle is α: r1=a1*b1 / [a1 2 *(sin(θ1-α)) 2 +b1 2 *(cos(θ1-α)) 2 ] 1 / 2 ; Wherein, a1 is the length of the major semi-axis of the elliptic curve corresponding to the first cross-sectional profile, b1 is the length of the minor semi-axis of the elliptic curve corresponding to the first cross-sectional profile, and θ1 is the angle between the ray extending from the center of the elliptic curve corresponding to the first cross-sectional profile to point m and the horizontal plane; Calculate the distance r2 from point n on the second cross-section profile to the center of the elliptical curve when the inclination angle is β: r2=a2*b2 / [a2 2 *(sin(θ2-β)) 2 +b2 2 *(cos(θ2-β)) 2 ] 1 / 2 ; Wherein, a2 is the length of the major semi-axis of the elliptic curve corresponding to the second cross-sectional profile, b2 is the length of the minor semi-axis of the elliptic curve corresponding to the second cross-sectional profile, and θ2 is the angle between the ray extending from the center of the elliptic curve corresponding to the second cross-sectional profile to point n and the horizontal plane; Calculate the misalignment Δr: Δr=|r1-r2|; Among them, in the extension direction of the shield tunnel, point m and point n coincide.

[0065] It should be noted that the inclination angle α and the inclination angle β in the above content may be equal or different.

[0066] For example, the misalignment analysis is conducted on the 28990.454 mileage ring gap. Figure 7 As shown, Figure 7 The cross-section extraction diagrams on both sides of the annular seam, where the red fitting curve represents the cross-section 6 cm to the left of the annular seam, and the cyan fitting curve represents the cross-section 6 cm to the right of the annular seam. Figure 8 The two sets of drawings are combined with the cross-section on both sides of the annular seam. The misalignment calculation method is used to calculate the misalignment in the 360-degree azimuth angle, as shown in the following example: Figure 9 According to the regulations on shield tunnel construction and acceptance, the limit of misalignment on both sides of the ring joint of subway tunnel is ±15mm. Figure 9 The red straight dotted line in the middle represents the limit value, and the range of the track plate is between 235° and 305°. It can be seen that the maximum misalignment at mileage 28990.454 is -9mm, which is less than the requirements specified in the specification.

[0067] The feasibility of the shield tunnel section misalignment detection method is verified below. A shield tunnel section is selected as an example for analysis. The tunnel is a circular segment shield tunnel with an inner diameter of 5.4m, an outer diameter of 6m, and a width of 1.5m per ring. A 3D laser scanner is used to scan the 1.5km section of the tunnel. Some of the original point cloud data are as follows: Figure 10 As shown in the figure, due to the large amount of data, in order to facilitate analysis and demonstration, 150m (100 rings) tunnel point cloud data is selected to calculate the misalignment for example verification.

[0068] First, assign serial numbers to each ring of segments. Since 100 rings of point cloud data are used this time, the corresponding ring numbers are 61-160. According to the requirements of the "Shield Tunnel Construction and Acceptance Specifications", the misalignment between ring joints should be controlled within 15mm. First, convert the three-dimensional point cloud data into a two-dimensional grayscale image, and then quickly detect the position of the ring joint. Then, perform cross-sectional slicing of 6cm on both sides of the corresponding ring joint mileage in the three-dimensional point cloud data. Use the shield tunnel section fitting method described in the first aspect to fit the two sections to an ellipse, and fit the two fitted ellipses. Finally, complete the calculation of the misalignment, and automatically output the maximum misalignment of each ring to form a Figure 11 The maximum misalignment distribution diagram between the rings is shown in the figure. Figure 11 The red dashed line in the middle is a curve with a limit of 15 mm. If the limit exceeds this line, the misalignment exceeds the specification requirements. The section mileage, ring number, start and end angles, arc length, and maximum misalignment of the sections on both sides of the longitudinal seam where the misalignment exceeds the specification requirements are recorded and formed into Table 2.

[0069] Table 2 Adjacent ring misalignment detection report Segment ring number Inter-ring gap left slice mileage Mileage of the right slice of the inter-ring gap Start and end angles (°) Arc length (m) Maximum misalignment during automatic detection (mm) Maximum misalignment after manual review (mm) 62~63 28350.454 28350.574 207~228 0.9896 19 17 63-64 28351.956 28352.076 2~31 1.3666 18 16 64-65 28353.453 28353.573 181~209 1.319 21 21 84~85 28368.449 28368.569 5~38 1.5551 18 16 90~91 28377.452 28377.172 201~213 0.5655 20 23 98~99 28389.454 28389.574 13~46 1.5551 21 24 108~109 28404.456 28404.576 200~225 1.1781 18 21 118~119 28419.452 28419.572 10~27 0.8011 16 13 133~134 28441.955 28442.075 158~176 0.8482 20 16 156~157 28476.458 28476.578 171~203 1.5080 17 15 The data for the 100 shield tunnel rings described above were manually reviewed, and the maximum misalignment of each ring was calculated. All rings with misalignment less than 15 mm calculated by 3D laser scanning were manually reviewed and found to meet the specification requirements. Ten segments with maximum misalignment greater than 15 mm were also manually reviewed in Table 2. The review results are shown in the column "Maximum misalignment after manual review" in Table 2. Manual re-measurement revealed that nine segments had maximum misalignment greater than 15 mm, and another segment, ring numbers 118-119, had a misalignment of 13 mm, while the misalignment automatically calculated by 3D laser scanning was 16 mm. The difference in misalignment between the two calculations was only 3 mm, and the accuracy of automatic misalignment detection by 3D laser scanning reached 90%.

[0070] The difference between the maximum misalignment of the manually measured 100 ring data and the misalignment calculated by 3D laser scanning was statistically analyzed. The statistical results are shown in Table 3, where the statistical quantity refers to the number of ring seams.

[0071] Table 3 Statistics of the difference in platform misalignment Misalignment difference (mm) Statistics Percentage (%) Cumulative percentage (%) (0~1] 10 10 10 (1~2] 19 19 29 (2~3] 51 52 81 (3~4] 11 11 92 (4~5] 8 8 100 As can be seen from Table 3, the overall difference in misalignment of the 3D laser scanning is controlled within 5 mm, of which the accuracy within 3 mm is 81%, the accuracy within 4 mm is 92%, and the accuracy within 5 mm is 100%.

[0072] In a third aspect, the present application provides a shield tunnel section misalignment detection system, comprising a three-dimensional laser scanning device and a computer device; The three-dimensional laser scanning device is used to scan the shield tunnel to obtain three-dimensional point cloud data of the shield tunnel, and is used to send the three-dimensional point cloud data to the computer device; The computer device includes a storage module, a processor, and a first computer program and a second computer program stored on the storage module and executable on the processor, wherein when the processor executes the first computer program, any one of the shield tunnel section fitting methods described in the first aspect is implemented, and when the processor executes the second computer program, any one of the shield tunnel section misalignment detection methods described in the second aspect is implemented.

[0073] The above is an explanation of the implementation mode of the present application by specific specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Although the description of the present application will be introduced in conjunction with some embodiments, this does not mean that the features of this application are limited to the implementation mode. On the contrary, the purpose of introducing the application in conjunction with the implementation mode is to cover other options or modifications that may be extended based on the claims of the present application. In order to provide an in-depth understanding of the present application, the above description contains many specific details. The present application can also be implemented without using these details. In addition, in order to avoid confusion or blurring the focus of the present application, some specific details will be omitted in the description. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other unless there is a conflict.

[0074] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A shield tunnel section fitting method, characterized in that: The following steps are involved: a. Obtain the 3D cross-section point cloud data of the shield tunnel; b. randomly extracting at least 5 three-dimensional cross-section point cloud data from the three-dimensional cross-section point cloud data and fitting an elliptic curve based on the extracted three-dimensional cross-section point cloud data; c. Calculate the deviation weight of each three-dimensional cross-section point cloud relative to the elliptic curve, and count the number of first point clouds whose deviation weights meet a first threshold, wherein the deviation weight is used to describe the degree of deviation of each cross-section point cloud relative to the fitted elliptic curve; d. If the number of the first point cloud meets the second threshold, output the elliptic curve; otherwise, the method returns to step b and iterates.

2. The shield tunnel section fitting method according to claim 1, characterized in that: The deviation weight is obtained by: Solve the coordinates of the two foci and the length of the major axis of the elliptic curve; Calculate the sum of the distances between the shield tunnel cross-section point cloud and the two foci; The absolute value of the major axis length, spacing and difference is used as the deviation weight.

3. The shield tunnel section fitting method according to claim 1, characterized in that: Step d further comprises: Before outputting the elliptic curve, execute step b and step c again to obtain the second point cloud quantity; If the number of the second point cloud is greater than the number of the first point cloud, the elliptic curve is updated.

4. The shield tunnel section fitting method according to claim 1, characterized in that: Step b further comprises: Mark the two intersection points of the tunnel profile with the tunnel floor; With the shield tunnel central axis as the center, construct the major arc between the two intersection points; Mark the quintiles of the major arc and construct rays from the median axis to the quintiles to obtain five sector-shaped areas; A 3D cross-section point cloud data is randomly selected in each of the five sector areas; The least squares method is used to perform ellipse fitting on the selected three-dimensional cross-section point cloud data.

5. A method for detecting the misalignment of a shield tunnel section, based on the shield tunnel section fitting method according to any one of claims 1 to 4, characterized in that: The following steps are involved: e. Obtain the three-dimensional point cloud data of the shield tunnel and obtain the annular seam position based on it; f. Along the extension direction of the shield tunnel, the first section and the second section are obtained before and after the annular seam; g. Using a shield tunnel section fitting method to obtain a first section profile and a second section profile of the first section and the second section respectively; h. Fit the first section profile and the second section profile and calculate the misalignment.

6. The method for detecting the cross-section misalignment of a shield tunnel according to claim 5, characterized in that: Step e includes: Map the shield tunnel 3D point cloud data into a 2D grayscale image; Locate the annular seam contour through a two-dimensional grayscale image; The Hough transform method is used to fit the annular seam and filter out noise points; The position of the annular seam is obtained based on the mapping relationship between the three-dimensional point cloud data and the two-dimensional grayscale image pixels.

7. The method for detecting the cross-section misalignment of a shield tunnel according to claim 6, characterized in that: Mapping shield tunnel 3D point cloud data into 2D grayscale images includes: The plane point cloud is rasterized using a 3-8 mm grid.

8. The method for detecting the cross-section misalignment of a shield tunnel according to claim 7, characterized in that: The plane point cloud is rasterized using a 5 mm grid.

9. The method for detecting the cross-section misalignment of a shield tunnel according to claim 5, characterized in that: The misalignment is calculated by: Calculate the distance r1 from point m on the first cross-section profile to the center of the elliptic curve when the inclination angle is α: r1=a1*b1 / [a1 2 *(sin(θ1-α)) 2 +b1 2 *(cos(θ1-α)) 2 ] 1 / 2 ; Wherein, a1 is the length of the major semi-axis of the elliptic curve corresponding to the first cross-sectional profile, b1 is the length of the minor semi-axis of the elliptic curve corresponding to the first cross-sectional profile, and θ1 is the angle between the ray extending from the center of the elliptic curve corresponding to the first cross-sectional profile to point m and the horizontal plane; Calculate the distance r2 from the n points on the second cross-sectional profile to the center of the elliptic curve when the inclination angle is β: r2=a2*b2 / [a2 2 *(sin(θ2-β)) 2 +b2 2 *(cos(θ2-β)) 2 ] 1 / 2 ; Wherein, a2 is the length of the major semi-axis of the elliptic curve corresponding to the second cross-sectional profile, b2 is the length of the minor semi-axis of the elliptic curve corresponding to the second cross-sectional profile, and θ2 is the angle between the ray extending from the center of the elliptic curve corresponding to the second cross-sectional profile to point n and the horizontal plane; Calculate the misalignment Δr: Δr=|r1-r2|; Among them, in the extension direction of the shield tunnel, point m and point n coincide.

10. A shield tunnel section misalignment detection system, characterized in that: Including 3D laser scanning equipment and computer equipment; The three-dimensional laser scanning device is used to scan the shield tunnel to obtain three-dimensional point cloud data of the shield tunnel, and is used to send the three-dimensional point cloud data to the computer device; The computer device includes a storage module, a processor, and a first computer program and a second computer program stored in the storage module and executable on the processor, wherein when the processor executes the first computer program, the shield tunnel section fitting method as described in any one of claims 1 to 4 is implemented, and when the processor executes the second computer program, the shield tunnel section misalignment detection method as described in any one of claims 5 to 9 is implemented.

Citation Information

Patent Citations

  • Circular shield tunnel section fitting method based on three-dimensional laser scanning technology

    CN111024037A

  • Tunnel point cloud data processing method and device and storage medium

    CN111813775A

  • Shield tunnel segment dislocation extraction method

    CN113091639A

  • Shield tunnel diameter convergence and radial slab staggering detection method based on laser scanning

    CN114370828A

  • Horseshoe-shaped mountain tunnel deformation detection method and system based on laser point cloud segmentation

    CN116007523A

Cited By

  • Tunnel segment ring ovality analysis method based on point cloud data and computer device

    CN119358102A

  • Shield segment slab staggering detection method and system fusing visual and geometric features

    CN120609285A

  • Shield segment misalignment detection method and system fusing visual and geometric features

    CN120609285B